Klasifikasi Jenis Pohon Menggunakan Convolutional Neural Network (CNN))

Pangaribuan, Salomo J. A. (2026) Klasifikasi Jenis Pohon Menggunakan Convolutional Neural Network (CNN)). Other thesis, Institut Teknologi Sepuluh Nopember.

Text (Salomo J A Pangaribuan Undergraduate Thesis) 07211940000051-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Di tengah krisis perubahan iklim, perdagangan karbon adalah salah satu langkah \mbox{mitigasi} yang dilakukan oleh pemerintah global. Salah satu subjek dari perdagangan karbon yang sangat vital adalah hutan. Salah satu tantangan dari perdangan karbon adalah adalah kebutuhan menghitung simpanan karbon pada hutan. Simpanan karbon dapat dihitung dengan menhitung nilai(\textit{Aboveground Biomass}) dari pohon pohon dihutan, yang dalam praktikiknya sangat bergantung pada pengukuran \textit{Diameter at Breast Height} (DBH). Sayangnya, selama ini pengukuran DBH masih dilakukan secara manual sehingga kurang efisien dan memakan banyak waktu. Penelitian ini hadir untuk menjawab tantangan tersebut dengan mengembangkan sistem otomatisasi berbasis \textit{Computer Vision} dan \textit{Deep Learning}. Dengan menggunakan algoritma YOLOv11s-Seg, sistem ini mampu melakukan \textit{instance segmentation} sekaligus mengklasifikasikan batang pohon dari 17 spesies berbeda secara otomatis. Melalui pengujian menggunakan \textit{dataset} citra \textit{smartphone} yang mencakup berbagai kondisi lapangan, model ini menunjukkan performa yang sangat impresif pada \textit{validation set}. Hasil evaluasi mencatatkan nilai \textit{mean Average Precision} (mAP) sebesar 0,9563 pada metrik mAP50-95(M), tingkat presisi masker mencapai 98,83\%, serta \textit{F1-score} sebesar 0,9729. Pencapaian ini membuktikan bahwa metode yang dikembangkan tidak hanya berbiaya rendah dan mudah digunakan, tetapi juga sangat layak menjadi solusi andal dalam mendukung sistem \textit{Measurement, Reporting, and Verification} (MRV) untuk perdagangan karbon yang lebih kredibel di Indonesia.
=================================================================================================================================
In the midst of the climate change crisis, carbon trading has emerged as a key mitigation strategy implemented by governments worldwide. Forests represent a vital component within the carbon trading framework. However, a significant challenge in carbon trading lies in the necessity to accurately quantify forest carbon stocks. Carbon storage is typically estimated by calculating the Aboveground Biomass (AGB) of trees, which in practice relies heavily on Diameter at Breast Height (DBH) measurements. Traditionally, DBH measurement has been conducted manually, a process that is often inefficient and time-consuming. This research addresses these challenges by developing an automated system based on Computer Vision and Deep Learning. Utilizing the YOLOv11s-Seg algorithm, the system is capable of performing instance segmentation while simultaneously classifying tree trunks across 17 different species automatically. Based on testing conducted with a smartphone imagery dataset encompassing various field conditions, the model demonstrated impressive performance on the validation set. Evaluation results yielded a mean Average Precision (mAP) of 0.9563 on the mAP50-95(M) metric, a mask precision rate of 98.83%, and an F1-score of 0.9729. These achievements prove that the developed method is not only cost-effective and user-friendly but also highly viable as a reliable solution to support Measurement, reporting, and Verification (MRV) systems for more credible carbon trading in Indonesia.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Perubahan Iklim, Perdagangan Karbon, Aboveground Biomass, Deep Learning, You Only Live Once (YOLO), Instance Segmentation.
Subjects:	T Technology > T Technology (General) > T57.5 Data Processing T Technology > T Technology (General) > T58.5 Information technology. IT--Auditing T Technology > T Technology (General) > T58.62 Decision support systems T Technology > T Technology (General) > T58.8 Productivity. Efficiency
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Computer Engineering > 90243-(S1) Undergraduate Thesis
Depositing User:	Salomo Ja Pangaribuan
Date Deposited:	20 Jan 2026 04:25
Last Modified:	20 Jan 2026 04:25
URI:	http://repository.its.ac.id/id/eprint/129790

Actions (login required)

View Item